Skip to main content

In the Weights

Category: AI Search / Retrieval

An AI-centric vanity search website that queries multiple large language models to score how well a person's name is 'remembered' in their training data weights. In the Weights was founded in 2026. The company is led by Thomas Dimson. Based in San Francisco, California, United States. Team size: 1-10. Latest round: Bootstrapped.

Founded
2026
Headquarters
San Francisco, California, United States
Team size
1-10

Value proposition

Provides a novel way to measure one's digital footprint in AI training data, shifting the paradigm from Google vanity searches to LLM-based personal information retrieval. Reveals which AI models 'remember' a person purely from their training weights without web search augmentation.

Products and solutions

In the Weights website — a vanity search tool that queries multiple AI models (Grok, Gemini, GPT-4.5/5/5.5, Claude, Llama, Qwen3, and others) to determine how strongly a person's name is encoded in the model's training data. Returns a strength score (0-996), shows which models recognize the person, highlights potential hallucinations, and features a public leaderboard.

Unique value

The first tool to quantify and compare how deeply individuals are encoded in the 'weights' of frontier AI models, offering a retro Nintendo-inspired interface and a public leaderboard of the most 'remembered' names across multiple LLMs.

Target customer

General public curious about their AI footprint; AI researchers and enthusiasts; media personalities and public figures interested in their LLM recognition score

Industries served

Consumer internet, AI awareness/education, Personal branding

Technology advantage

Queries multiple frontier LLMs (Grok, Gemini, GPT-4.5/5/5.5, Claude, Llama, Qwen3) simultaneously, clusters similar descriptions, and assigns a composite strength score. Flags model hallucinations. Built by former OpenAI employees with deep expertise in AI systems and recommendation algorithms.

How they differentiate

Unlike traditional Google vanity searches that index web pages, In the Weights probes the internal knowledge stored in LLM weights — measuring what AI models 'know' about a person without web search. Provides cross-model comparison, hallucination detection, and a gamified leaderboard.

Main competitors

Google search (traditional vanity search), ChatGPT/Claude/Gemini direct queries (manual single-model checks), AI model leaderboards like Chatbot Arena (different focus — model comparison vs. personal recognition)

Key partnerships

Queries models from OpenAI (GPT series), Anthropic (Claude), Google (Gemini), xAI (Grok), Meta (Llama), Alibaba (Qwen3), and other LLM providers

Major milestones

Launched June 18-20, 2026, Featured on TechCrunch (June 20, 2026), Reached #1 on Product Hunt daily leaderboard (June 20, 2026), Viral social media traction with hundreds of thousands of views on X/Twitter

Growth metrics

Viral launch with 331K+ views on Thomas Dimson's X announcement post; featured on TechCrunch, The Decoder, Digg, Product Hunt (#1 daily product)

Market positioning

Novel side project / viral consumer tool occupying a new niche — 'AI vanity search' — at the intersection of personal branding, AI literacy, and model evaluation. Not positioned as a commercial product but as a cultural curiosity that struck a nerve.

Geographic focus

Global (English-language focused)

About Thomas Dimson

Ex-Member of Technical Staff at OpenAI (worked on ChatGPT, post-training, Sora); Ex-CEO & Co-founder of Global Illumination (acquired by OpenAI); Ex-Director of Engineering at Instagram (authored Instagram's content ranking algorithms for feed, stories, and Explore); Ex-Software Engineer at Amazon; Stanford University

Official website: